基于磁敏感血管征与脑CT灌注参数对急性前循环大血管闭塞性卒中静脉溶栓后出血转化的预测价值

Value of susceptibility vessel sign and brain computed tomography perfusion parameters in predicting hemorrhagic transformation after intravenous thrombolysis in acute anterior circulation large vessel occlusion stroke

摘要:

目的 探讨磁敏感血管征联合脑计算机体层成像（CT）灌注参数对急性前循环大血管闭塞性卒中（LVOS）静脉溶栓后出血转化（HT）的预测价值。方法 选取2022年2月至2025年1月在珠海市第五人民医院进行静脉溶栓治疗的急性前循环LVOS患者160例，根据溶栓后24 h是否发生HT分为HT组（46例）与非HT组（114例）。收集临床资料、磁敏感血管征及脑CT灌注参数［脑血流量（CBF）、脑血容量（CBV）、流量提取乘积（FED）、表面通透性（PS）、达峰时间（TTP）］。采用多因素Logistic回归分析评估静脉溶栓后HT的影响因素，构建联合预测模型，采用受试者操作特征（ROC）曲线及曲线下面积（AUC）评估预测模型的效能。结果 HT组入院美国国立卫生研究院卒中量表（NIHSS）评分、磁敏感血管征阳性率、FED、PS、TTP均高于非HT组（P<0.05），CBF、CBV低于非HT组（P<0.05）。多因素Logistic回归显示，磁敏感血管征阳性（OR=2.319）、FED升高（OR=4.778）、PS升高（OR=1.765）、TTP延长（OR=1.982）是静脉溶栓后HT的危险因素，CBF（OR=0.535）、CBV（OR=0.618）为保护因素。ROC曲线显示联合模型AUC为0.932，灵敏度为93.86%，特异度为84.78%。结论 基于磁敏感血管征和脑CT灌注参数的联合模型对LVOS患者静脉溶栓后HT具有较高预测价值，有助于早期识别高危患者，并指导个体化溶栓决策。

Abstract:

Objective To investigate the value of susceptibility vessel sign （SVS） combined with brain computed tomography （CT） perfusion parameters in predicting hemorrhagic transformation （HT） after intravenous thrombolysis in acute anterior circulation large vessel occlusion stroke （LVOS）.Methods A total of 160 LVOS patients who underwent intravenous thrombolysis in Zhuhai Fifth People's Hospital from February 2022 to January 2025 were enrolled， and according to the presence or absence of HT at 24 hours after thrombolysis， the patients were divided into HT group with 46 patients and non-HT group with 114 patients. Related data were collected， including clinical data， SVS， and brain CT perfusion parameters ［cerebral blood flow （CBF）， cerebral blood volume （CBV）， flow extraction product （FED）， permeability surface （PS）， and time to peak （TTP）］. The multivariate logistic regression analysis was used to identify the influencing factors for HT after intravenous thrombolysis， and a combined predictive model was established. The receiver operating characteristic （ROC） curve and area under the curve （AUC） was used to assess the predictive performance of the model.Results Compared with the non-HT group， the HT group had significantly higher National Institutes of Health Stroke Scale score on admission， SVS-positive rate， FED， PS， and TTP and significantly lower CBF and CBV （all P<0.05）. The multivariate logistic regression analysis showed that positive SVS （OR=2.319）， elevated FED （OR=4.778）， elevated PS （OR=1.765）， and prolonged TTP （OR=1.982） were independent risk factors for HT， while CBF （OR=0.535） and CBV （OR=0.618） were protective factors. The ROC curve analysis showed that the combined model had AUC of 0.932， a sensitivity of 93.86%， a specificity of 84.78%.Conclusions The combined model based on SVS and brain CT perfusion parameters has a high value in predicting HT after intravenous thrombolysis in LVOS patients， which may help to identify high-risk patients in the early stage and guide decision-making for individualized thrombolysis.